U.S. patent application number 14/066307 was filed with the patent office on 2014-02-20 for method for monitoring demagnetization.
The applicant listed for this patent is Pedro Rodriguez. Invention is credited to Pedro Rodriguez.
Application Number | 20140049285 14/066307 |
Document ID | / |
Family ID | 46026794 |
Filed Date | 2014-02-20 |
United States Patent
Application |
20140049285 |
Kind Code |
A1 |
Rodriguez; Pedro |
February 20, 2014 |
Method For Monitoring Demagnetization
Abstract
A method for discovering demagnetisation faults of a permanent
magnet synchronous generator, such as a wind power generator. The
method is performed during operation of the synchronous generator
and includes measuring the vibration of the stator, performing a
frequency analysis of the vibration, and deducing whether the
generator suffers from demagnetization of a permanent magnet, from
the vibration analysis. Moreover, geometric eccentricity faults and
electric short circuit faults may also be detected from the
vibration.
Inventors: |
Rodriguez; Pedro; (Vasteras,
SE) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Rodriguez; Pedro |
Vasteras |
|
SE |
|
|
Family ID: |
46026794 |
Appl. No.: |
14/066307 |
Filed: |
October 29, 2013 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
PCT/EP2012/057555 |
Apr 25, 2012 |
|
|
|
14066307 |
|
|
|
|
Current U.S.
Class: |
324/765.01 |
Current CPC
Class: |
G01R 31/343 20130101;
G01H 1/003 20130101; H02P 29/02 20130101 |
Class at
Publication: |
324/765.01 |
International
Class: |
G01R 31/34 20060101
G01R031/34 |
Foreign Application Data
Date |
Code |
Application Number |
Apr 29, 2011 |
EP |
11164289.8 |
Claims
1. A method for monitoring demagnetization of permanent magnets in
a synchronous machine, such as a wind power generator, having a
stator with windings and a rotor with permanent magnets being
arranged to rotate in relation to the stator, the method being
performed during operation of the synchronous machine, and the
method including the steps of: measuring the vibration of the
stator, performing a frequency analysis of the vibration,
determining, on the basis of whether the magnitude of the vibration
at the supply frequency of the stator exceeds a threshold value or
not, whether the machine suffers from demagnetization of a
permanent magnet or not.
2. The method according to claim 1, wherein the method further
comprises the step of: determining, on the basis of whether the
magnitude of the vibration at the rotation frequency exceeds a
threshold value or not, whether the machine suffers from
demagnetization of a permanent magnet or not.
3. The method according to claim 1, wherein the vibration
measurement includes detecting vibrations by means of a vibration
sensor that is fixed to a frame of the stator or fixed to the
stator.
4. The method according to claim 1, wherein the machine is a
permanent magnet synchronous generator.
Description
FIELD OF THE INVENTION
[0001] The invention relates to systems for monitoring a permanent
magnet synchronous machine, such as a generator. Especially it
concerns monitoring magnetic faults of the permanent magnets of a
synchronous generator, for example in a permanent magnet
synchronous generator for wind power generation.
BACKGROUND OF THE INVENTION
[0002] An important factor enabling high reliability at electric
power plants is to provide generators, such as wind generators,
with condition monitoring systems in order to detect faults at an
early stage. The invention aims to provide an improved diagnosing
method for detecting and identifying magnetic faults of permanent
magnet synchronous generators (PMSG), especially for wind power
permanent magnet synchronous generators. Permanent magnets
synchronous generators are one of the common machines in the wind
generation industry. Detecting demagnetization is important since
it produces degradation of the machine performance. There have been
methods for detecting demagnetization based on monitoring the back
EMF (electromotive force) and the stator current of the
generators.
[0003] However, when monitoring those indicators, there is a risk
that wrong information could be given since other faulty conditions
produce similar signatures than demagnetization.
[0004] U.S. Pat. No. 7,324,008 (D1) describes analyzing an
electrical machine using finite element method (FEM) analysis with
at least one fault condition to be able to predict the effect of
the fault condition. The result of the FEM analysis can be used to
identify the analyzed faults from live measurements of the machine
(see D1, abstract). D1 describes a transverse flux motor but also
suggests that similar FEM analysis with a fault could be used for
other electrical machines (see D1, column 7, line 26-50). D1
suggests simulating magnetic faults and the effect of the magnetic
faults to the magnetic flux, so that measurements of the magnetic
flux, by means of "search coils", can be used as an indication of
magnetic faults (see D1, column 1, lines 46-51). The method
described in D1 provides means for detecting faults that degrade
the magnetic strength of magnets due to overheating and/or
demagnetization (see D1, column 6, line 15-16).
[0005] In the technical field it is also important to monitor the
rotor bearings. The condition of the rotor bearings of generators
for wind turbines are often monitored by measuring the vibration
close to the bearings.
[0006] D1 suggests analyzing other fault conditions, such as
mechanical and electrical misalignments, and suggests also using
other sensors for monitoring the machine, in addition to the coils
used for sensing the magnetic flux (see D1, column 7, line 27-31),
such other sensors as temperature, vibration and current
sensors.
[0007] The invention concerns diagnosis of permanent magnet
synchronous machines, especially detecting magnetic faults of
generators, and provides an alternative to using search coils.
[0008] The article "Static-, Dynamic-, and Mixed-Eccentricity Fault
Diagnoses in Permanent-Magnet Synchronous Motors", Ebrahimi et al,
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 56, NO. 11,
NOVEMBER 2009 4727, (D4) indicates how static-, dynamic- and
mixed-eccentricities influences the current frequency spectra with
amplitudes in sideband components of the stator current frequency.
Also, D4 describes how the current spectra are influenced by
demagnetisation, short circuit faults and open circuit faults, and
can be distinguished from the eccentricity faults by not creating
sideband components. The invention provides an alternative to the
current spectrum analysis for discovering demagnetisation, as
described in D4, but can also be used in addition to such stator
current analysis.
[0009] D1 and D4 are considered the most relevant prior art
documents, but reference will in the following be made to a few
documents concerning solving other technical problems than
discovering demagnetization of permanent magnets in synchronous
machines.
[0010] US2011/0018727 (D2) describes a method and device for
performing wind turbine generator fault diagnostics, wherein
sensors monitor a wind generator and signals from the sensors are
analyzed to detect anomalies that indicate faults. The method that
is described in D2 evaluates (see D2, FIG. 1) electrical signals
(voltage, current), vibration signals and temperature signals. The
electrical signals and vibration signals are subjected to
respective spectral analysis (see D2, references 110 and 120) (such
as an FFT or Fast Fourier Transform). The spectra are subjected to
signature analysis (see D2, references 140 and 142) and anomaly
detection (see D2, references 150 and 154). Transients in the
temperature are detected and anomalies of the temperature
identified (see D2, reference 156). Upon detecting anomalies in the
electric and vibration spectra, respectively, and in the
temperature, it is concluded that the generator system has
electrical or mechanical faults (see D2, paragraph [0009]).
Vibration signals from an accelerometer and a voltage signature
determined from the generator voltage signals may be used to detect
an underlying eccentricity fault (see D2, paragraph [0010]). The
document describes generators in a general way and does not
describe detection of magnetic faults, and especially not detecting
magnetic faults in permanent magnet synchronous generators.
[0011] Analysis of the vibration signals has been used to diagnose
faults of electric machines, see Jover Rodriguez, P. V., 2007,
"Current-, Force-, and Vibration-Based Techniques for Induction
Motor Condition Monitoring", Doctoral Dissertation, Helsinki
University of Technology, Finland (D3), which may be found at
[0012]
http://lib.tkk.fi/Diss/2007/isbn9789512289387/isbn9789512289387.pdf
[0013] D3 describes analyzing the frequency spectra of stator
vibrations in an asynchronous motor. The aim of this research was
to discover the best indicators of induction motor faults, as well
as suitable techniques for monitoring the condition of induction
motors. D3 describes the effects of electromagnetic force on the
vibration pattern when the motor is working under fault conditions.
Moreover, D3 describes a method that allows the prediction of the
effect of the electromechanical faults in the force distribution
and vibration pattern of the induction machines. In D3, FEM
computations are utilized, which show the force distributions
acting on the stator of the electrical machine when it is working
under an electrical fault. It is shown that these force components
are able to produce forced vibration in the stator of the machine.
The results were supported by vibration measurements. The
low-frequency components could constitute the primary indicator in
a condition monitoring system. D3 uses vibrational analysis to
detect faults of an induction motor, which faults are broken rotor
bars, broken end ring, inter-turn short circuit, bearing and
eccentricity failures.
SUMMARY OF THE INVENTION
[0014] An aim of the invention is to provide a method for detecting
magnetic problems, which is easy to use, still being reliable. The
present invention provides a method for detecting magnetic faults
of a permanent magnet synchronous machine comprising a stator with
windings and terminals for the machine currents, and a rotor
provided with permanent magnets, which rotor is rotatably arranged
within the stator. By detecting and identifying magnetic faults,
especially demagnetization of a permanent magnet, reparation of a
machine is facilitated, since exchanging the permanent magnets of
the rotor will remove the malfunction and improve the performance.
For these purposes, the present invention provides a method for
monitoring demagnetization of permanent magnets in a synchronous
machine. The method is performed during operation of the
synchronous machine. The method includes measuring the vibration of
the stator, performing a frequency analysis of the vibration,
determining, on the basis of whether the magnitude of the vibration
at the supply frequency (f.sub.S) of the stator exceeds a threshold
value or not, whether the machine suffers from demagnetization of a
permanent magnet or not.
[0015] Although, this application mainly describes monitoring and
fault analysing of permanent magnet synchronous generators, the
invention is also beneficial for monitoring and fault analysing of
permanent magnet synchronous motors.
[0016] According to one embodiment of the invention, the method
further comprises the step of determining, on the basis of whether
the magnitude of the vibration at the rotation frequency (f.sub.R)
exceeds a threshold value or not, whether the machine suffers from
demagnetization of a permanent magnet or not.
[0017] Demagnetization of a permanent magnet in a synchronous
machine affects the vibration spectrum at the stator supply
frequency (f.sub.S) and at the rotation frequency (f.sub.R) of the
rotor. Monitoring the vibration at the stator supply frequency
(f.sub.S) can be used to discover a magnetic fault, and distinguish
the magnetic fault from other faults, such as electric and
geometric faults. Monitoring also the vibration magnitude at the
rotation frequency makes the deduction of a magnetic fault
clearer.
[0018] Measuring the vibration provides an easy way to monitor
demagnetization. The invention makes it possible to deduce that a
permanent magnet synchronous machine suffers from demagnetization
of a permanent magnet from the vibration alone. The vibration
measurement may include detecting vibrations by means of a
vibration sensor that is fixed to a frame of the stator or fixed to
the stator. For example, the inventive method can be realized by
the following steps; positioning and fixing a vibration sensor,
such as an accelerometer or accelerometers, to the stator frame;
measuring the vibration of the stator frame; analyse the
frequencies of the vibration signal from the accelerometer;
identify magnitudes of frequencies indicating demagnetisation and
determine if the generator suffers from demagnetisation of a
permanent magnet.
[0019] Search coils or other electrical measurements are not
needed. Demagnetization faults can be identified from the magnitude
of the vibration at the supply frequency of the stator or at the
rotation frequency of the rotor.
[0020] The method can be implemented in monitoring equipment that
can be connected to a permanent magnet synchronous generator, e.g.
attaching an accelerometer to the stator frame or connecting an
already installed accelerometer to a computing device, provided
with diagnostic means, of the monitoring equipment.
[0021] Alternatively, the method can be implemented in a control
system of, for example, a wind power plant having wind turbines
comprising permanent magnet synchronous generators, which control
system receives vibration signals from vibration sensors fixated at
the respective stator frames of the generators, which method is
implemented in such a control system by analyzing the vibration
signals and detecting magnetic faults from the frequency spectra of
these vibration signals.
[0022] The invention provides a method that can be used to identify
demagnetisation, from vibrations alone. Current measurements can be
made in addition to the vibration sensing. In an embodiment, the
method includes measuring at least one current of the stator,
preferably each branch current, and performing a frequency analysis
for the current, such as for each branch current. Preferably, the
method includes monitoring frequencies of the current spectrum
indicating electric faults, magnetic faults and/or geometric
faults, to find indications of such faults in the current or
currents. In an embodiment, the system comprises means for
frequency analysis of a stator current and means for determining
faults based on the frequency analysis.
BRIEF DESCRIPTION OF THE DRAWINGS
[0023] FIG. 1 shows a method for detecting demagnetisation faults
in accordance with the invention.
[0024] FIG. 2 shows a system for detecting faults in accordance
with the invention.
[0025] FIG. 3 illustrates a vibration spectrum indicating a
demagnetisation fault.
DETAILED DESCRIPTION OF THE INVENTION
[0026] FIG. 1 illustrates a method for detecting magnetic faults on
a permanent magnet synchronous generator. The method starts by
obtaining 11 a vibration signal from a permanent magnet synchronous
generator. The obtained signal is subjected to a frequency analysis
12 wherein the vibration signal is divided into a spectrum of
frequencies and the magnitude of each frequency is calculated. The
frequency analysis 12 is suitably performed by an FFT-analysis
(Fast Fourier Transform) or the like for the stationary case, or,
for example, using a discrete wavelet transform during transients
due to the wind speed variations. The magnitudes in the spectrum of
vibration frequencies of a faulty generator may deviate from those
of a properly operating generator. The spectrum is therefore
subjected to a fault analysis (step 13), which monitor the
magnitudes of the vibration frequencies to discover faults in the
generator, especially a demagnetisation of a permanent magnet of
the generator. In more detail, the fault analysing step monitors
the magnitudes of the vibration at the stator supply frequency
(f.sub.S) and the rotation frequency (f.sub.R) of the rotor.
Heightened levels at these two frequencies in the vibration
spectrum indicate a deviation of magnetic level of a permanent
magnet in the rotor. The obtaining step 11 may include a plurality
of sub steps, for example, positioning and fixing a vibration or
motion sensor, such as one or more accelerometers, and possibly
also, as an alternative or in addition, electronic gyroscopes, to
the stator or a frame to which the stator is secured. The obtaining
step 11 may also include the substeps of measuring vibrations by
means of the sensor, and the substep of transferring the vibration
signal to fault analysing equipment, including receiving the
vibration signal in the fault analysing equipment.
[0027] Preferred embodiments of the present invention includes
detecting other faults in addition to the demagnetization faults,
which other faults are geometric faults, such as a static or
dynamic eccentricity, or a mixed eccentricity, and electric faults,
such as short circuit of stator windings. A similar FEM modelling
technique as used in D3 may be used to provide reliable diagnostic
methods for fault detection of the permanent magnet synchronous
generator in question.
[0028] The method of FIG. 1 illustrates obtaining 14 and analysing
15 the frequency of each stator branch current as optional steps
(as indicated by broken lines) to provide an enhanced basis for the
fault analysis 13. In such an embodiment, the fault analyser is
adapted to deduce magnetic, electric and geometric faults from both
the stator currents and the stator vibration.
[0029] The fault analysis 13 may be followed by subsequently
correcting the detected faults (step 16), such as substituting a
faulty permanent magnet, and/or adjusting other detected faults of
the generator.
[0030] FIG. 2 illustrates a vibration sensor 5 communicatively
connected to a fault analyser 21 with means for analysing vibration
signals to detect magnetic faults of a permanent magnet synchronous
generator. The vibration sensor is secured to a permanent magnet
synchronous generator 1 to sense the vibrations of a stator 2 of
the generator 1, and is arranged to obtain the vibration signals
for analysis upon rotation of the rotor 3 of the synchronous
generator 1. The fault analyser 21 comprises a vibration signal
interface 22, exemplified by a sensor interface 22 for receiving a
wired or wireless connection to a vibration sensor, such as an
accelerometer, and receiving vibration signals, a spectrum analyser
23, a fault identifier 24 and an output for an operator in the form
of a display 25. The spectrum analyser 23 is adapted to analyse the
frequency spectrum of a vibration signal received by the fault
analyser 21. For this purpose the spectrum analyser 23 applies a
Fourier transform or a time frequency decomposition to the received
signal, for example use a FFT (Fast Fourier Transform) or wavelet
transform and produces amplitude levels for the frequencies
constituting the vibration signal, so that the signatures of the
frequency components of the vibration can be identified. The fault
identifier 24 receives the frequency spectrum including the
magnitudes of each respective frequency from the spectrum analyser
23. The fault identifier 24 is adapted to identify the frequency
signatures of the vibration signal, and to link these signatures to
a specific fault condition, especially identifying frequencies
having magnitudes differing from a healthy generator and deducing
what type of fault the generator is suffering from. The fault
identifier 24 is especially adapted to identify demagnetisation of
the permanent magnets of the generator by monitoring the
frequencies of the vibration spectrum indicating a demagnetisation
fault and monitoring the magnitudes of these frequency components
of the spectrum. To be able to estimate the severity of a vibration
frequency indicating a demagnetisation fault, the fault identifier
suitably includes, or has access to a memory, with reference data,
such as magnitude levels corresponding to levels of
demagnetisation. This reference data can be created by measuring a
generator, or a motor, during operation with a faulty permanent
magnet, such as a permanent magnet less strong than the nominal
magnetic strength of the permanent magnets normally used in the
machine. For example, using a permanent magnet having 80 percent of
the strength of the permanent magnets of a generator during normal
operation, and measuring the vibration gives a measure of the size
of demagnetisation faults. The magnitudes of the vibration at a
demagnetisation fault indicating frequency can suitably be
interpolated and extrapolated from results from simulations and
also approximated to be proportional to the level of
demagnetisation. An alternative method to obtain reference data to
the fault identifier 24 of the fault analyser 21 is to perform a
FEM-modelling of the generator with different levels of
demagnetisation of the permanent magnets.
[0031] The signature for demagnetisation is seen in the vibration
frequency spectrum, especially at the rotation frequency (f.sub.R)
of the rotor and the stator supply frequency (f.sub.S) (see FIG.
3). The frequency component that gives the best indication of a
demagnetisation fault is the supply frequency (f.sub.S) of the
stator. In a healthy machine there is substantially no vibration at
this frequency i.e. the magnitude of the vibration at the supply
frequency (f.sub.S) of the stator is close to zero. Therefore, the
fault identifier 24 is adapted to compare the magnitude of the
vibration at the supply frequency (f.sub.S) of the stator with the
reference data. On the basis of the reference data a threshold
value for a fault condition can be determined. If the magnitude of
the vibration at the supply frequency (f.sub.S) of the stator
exceeds the threshold value, the fault identifier 24 determines
that the machine suffers from demagnetization of a permanent
magnet. Otherwise the fault identifier 24 determines that the
machine does not suffer from demagnetization of a permanent
magnet.
[0032] Even in a healthy machine there are typically vibrations at
the rotation frequency (f.sub.R) of the rotor, and consequently a
vibration at this frequency is not an optimal indicator of a
demagnetisation fault. However, vibration at the rotation frequency
(f.sub.R) can be used to confirm the fault identification results
obtained from the analysis at the supply frequency (f.sub.S) of the
stator. From the reference data a threshold value for the magnitude
of the vibration at the rotation frequency (f.sub.R) corresponding
to a fault condition can be determined, and the fault identifier 24
is adapted to compare the magnitude of the vibration at the
rotation frequency (f.sub.R) with the threshold value. The
determination of whether the machine suffers from demagnetization
of a permanent magnet or not is done in a corresponding manner as
when using the vibrations at the supply frequency (f.sub.S) of the
stator as the fault indicator.
[0033] The fault analyser 21 is suitably provided to present the
result of the analysis, especially an identified demagnetisation
fault, on a user interface in the form of a display 25. In
addition, or alternatively, an audible alarm or other fault
indication may be presented via a loudspeaker (not shown).
[0034] Apart from the sensor interface 22 for the vibration signal,
which may arrive via a computer network, the fault analyser 21 is
suitably provided with other sensor interfaces for interfacing
other sensors, such as a contact 26 for receiving measurement
signals from an additional sensor. Alternatively, the same
interface may be arranged for receiving measurements from different
measuring units. The fault analyser 21 exemplified in FIG. 2
includes a contact 26 for interfacing and receiving signals from
another sensor, especially a current meter, which current meter
(not illustrated) should be arranged for measuring the branch
currents of the generator 1. The current signals are fed to a
current analyser 27 provided to perform a frequency spectrum
analysis, such as an FFT-analysis, and the current analyser 27
subsequently transfers the frequency spectrum magnitudes to the
fault identifier 24. The fault identifier is adapted to identify
faults indicated by magnitudes in the, or each, stator branch
current spectrum, like electrical faults, e.g. inter-turn
short-circuits of a stator winding.
[0035] The fault analyser 21 may be integrated in a portable
service and control equipment, which may be connected to a
vibration sensor fixed to a stator frame and arranged to obtain
stator vibrations. The fault analyser 21 may alternatively be
integrated into a control and monitoring equipment permanently
arranged in a control room for monitoring and controlling the
generator, such as in a control room at a wind power plant.
[0036] The spectrum analyser 23 and the fault identifier 24 are
illustrated as separate entities, but can suitably be provided as a
combination of software and hardware entities in a computer and for
example share processor and memory. In the same computer, the
current analyser 27 may suitably be integrated.
[0037] The magnetic fault monitoring method may be implemented as a
computer program product, and include program steps to deduce
whether a machine has a demagnetization fault. When the computer
program is executed on a computer that receives vibration signals
from a permanent magnet synchronous machine as input, the computer
program should be adapted to discover whether the permanent magnet
synchronous machine suffers from a demagnetization fault, or not.
The program should be adapted to make technical considerations
based on the vibrations of the permanent magnet synchronous
machine, such as discovering a magnetic fault. Especially, the
program is adapted to link demagnetization faults to operational
parameters of the permanent magnet synchronous machine, such as the
rotation frequency f.sub.R and/or the stator supply frequency
f.sub.S. By doing this, the program solves the technical problem of
deducing whether a permanent magnet synchronous machine suffers
from a demagnetization fault or not, taking technical
characteristics of the machine into consideration, i.e. operational
frequencies of the machine, when analysing the physically obtained
vibrations of the machine. The program should suitably also be
adapted to provide an output of the result of the deduction to an
operator.
[0038] FIG. 3 illustrates a vibration spectrum of a permanent
magnet synchronous generator having a demagnetised permanent
magnet. The vibration magnitude at the stator supply frequency
f.sub.S, as well as the vibration magnitude at the rotation
frequency of the rotor f.sub.R is affected and both magnitudes are
higher than normal, especially the vibration magnitude at the
stator supply frequency f.sub.S. Similarly, signatures of geometric
and electric faults should suitably be identified in the frequency
spectrum of the vibration signal.
[0039] A system, method and program product for discovering
demagnetisation faults of a permanent magnet synchronous generator,
such as a wind power generator has been described. The method is
performed during operation of the synchronous generator and
includes measuring the vibration of the stator, performing a
frequency analysis of the vibration, and deducing, from the
vibration analysis, whether the generator suffers from
demagnetization of a permanent magnet.
* * * * *
References